“Contributors to Linux are nothing short of dedicated when it comes to offering their coding efforts, but as many are aware, much of that effort is wasted in the way of duplicated work, a great deal of which happen to only be the “sexy” parts of the code base. The problem is not so much in getting developer support, but in getting the masses organized and motivated to tackle the otherwise neglected aspects of the open source operating system.” The following osViews editorial contributor has some interesting ideas to not only help Linux development but also the platform as a whole and even its promotion.
Linux Home Desktop Kit PC Project: code named Gates Crusher
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
22 Comments
2004-01-06 5:14 pm
Anonymous
OT, is osViews runned by the same people that runned osOpinion? Strikingly similar styles, similar slogan, everything.
2004-01-06 5:19 pm
Anonymous
Unfortunately, the OSS community considers “GUI or not”
———–
Of course “GUI or not” is a design decision! There are a lot of Linux users who do not like tho use the GUI and don’t want to. They represent a market. The CLI players available represent someone serving that market. Supply and demand, that’s all it is.
“Gnome or KDE”
————-
Again, of course GNOME or KDE is a design decision. There is a reason why existing GNOME and KDE users get pissed when people talk about getting rid of one or the other. They are very different. A KDE app is tightly integrated into KDE and its framework. A GNOME app is more independent and is often better for those who use non DE at all, like Fluxbox or Window Maker. One goes for a very minimal UI, while the other avoids trading power for simplicity. Saying that writing seperate GNOME or KDE apps is a duplication of effort is like saying that Microsoft should stop development of Windows Media Player, because iTunes now runs on Windows. Or that Apple should stop development of the iApps because there are other programs that do the same thing. Its not a “one-size fits all” market. Some people demand applications done “the Apple way” or “the GNOME way” or “the KDE way” and people are supplying that demand.
I seem to miss all the fun stuff… fonts with *embedded* *comments*? What for, ye flippin’ gods? Ups, off-topic…
———-
Fonts can contain comments, just like code. Only a few do, and there was a problem with the FreeType parser not completely ignoring those comments. This is the kind of drudge work that gets done everyday.
Of course. But how many different ways of decoding an MP3 to line-out does one *need*?
——–
Well, do you want a player for stand-alone environments (do the decoding yourself) or one integrated into a media framework (depend on the framework). People in resource-constrained environments will balk at having to start up aRts, while people sitting on KDE will balk at an app that doesn’t use aRts. Then, do want a GUI player or a CLI player? Again, a lot of people are more comfortable with the CLI. Do you want a skinnable player that looks really flashy (WinAmp), or a lightweight, barebones player that gets the job done (Kaboodle). There are lots of possibilities.
How many ways of printing a document?
——–
One. *NIX apps almost invariably generate Postscript and send it to a printer driver. Usually, this is CUPS, which has become the standard on UNIX. kde-print and gnome-print are API-level details, and of no concern to the user.
How many ways of setting the volume of your sound output?
———
aRts or ALSA/OSS? Or Jack? Remember, different people have different needs. Pro-audio people need the low-latency of Jack. Desktop users want the tons of codecs provided by aRts. People who use a resource-constrained environment probably don’t want anything more than OSS or ALSA.
But it *is* a problem that the Linux user simply doesn’t start with a “clean” OS and picks his/her choice in tools.
——–
Why do you do an “install everything?” Just install the default desktop, or choose a minimal install. Last I checked, RedHat only installed one set of tools for each task — I presume other distros are similar.
They’re free, so they’re all on the CD / the repository, and end up on the user’s hard drive.
———
This is an interesting point. People need to experiment. Experienced Windows and Linux users often take it for granted that people know what apps are best for a given task. Consider someobody who wants to chat with friends. What client does she use? download.com lists 279 programs in its “chat” directory. How are they supposed to know that everyone uses AIM? Windows users come to Linux knowing things like “Word is the ‘standard’ word processor, AIM is the ‘standard’ IM client, etc” When they get to Linux, they don’t know that “Kopete is the ‘standard’ KDE IM client” or “GAIM is the ‘standard’ GNOME IM client” and are confused by all the available applications. That’s an unavoidable problem of changing OSs. Distros can do something to help, but perhaps the most helpful would be a tutorial site that would collect a few basic pieces of information from the user (desktop, power of machine, taste in features/power) and suggested a “standard” mix of software for a given task.
But the average Linux beginner is dropped right in the middle of a crowded “Start” menu,
———
They shouldn’t have chosen “Install Everything”! The default Fedora desktop comes pretty bare, so is the default Debian desktop.
with a /usr/bin too long to get an overview over
———
Only Windows users ever look in /usr/bin. Linux users know that there is nothing in there they need to know about, and avoid looking.
, and program names that don’t give a hint as to their purpose. IMHO, *that’s* the major flaw with Linux as it’s presented today.
———-
What does WinAmp mean? What does PowerPoint mean? What does Excel mean? Meanwhile, the KDE equivilents are KPlayer, KPresenter, and KSpread. I know people hate the ‘K’ convention, but it goes to show you that its very useful for making comprehensible program names.
2004-01-06 5:58 pm
Anonymous
Never mind, don’t need to answer that question earlier. osViews is runned but Kelly McNeil. Like on OSNews, he has the nerve to claim his site isn’t biased because it is community based – does that mean MacSlash isn’t biased because it is community based? He claims his site, in his own words, ” a technology barometer that gauges the overall opinions of the technology-focused community”. In other words, he is saying that his “community” represents fairly every section of the “technology-focused community” that it can be a barometer. Reading the past few articles, all I can say is “Hah! Yeah right” (sacarstically, of course).
Nothing against Kelly McNeil, but why does he keep on denying he has biases and thus his site woudl attract those with similar biases? Hiss choice of XServe as their server shows their bias.
2004-01-06 6:41 pm
2004-01-06 11:08 pm
Anonymous
I approve whole heartedly. I also wish he would come up with something us non-techie linux lovers could do. I’ve got 5-6 hours a week I could contribute.
2004-01-06 11:35 pm
Anonymous
All ppl new to GNU/Linux will probably make the same mistake of thinking “There’s a lot of duplicate efforts”. I also did.
It’s hard to understand that somebody who spends his time not inventing his wanted next generation editor will _not_ automatically spend it on wrinting an application which seems more needed (from the desktop users point of view).
However, requests like these will remains to appear unless one of the both big Desktops is “complete”. Right now none is. For example, a lot of KDE users will probably need The Gimp while a lot of Gnome users will need K3B.
As long as this situation remains, it’s hard to pack a distribution with only one enviroment which thus appears “consistent” for the usual home user. (Consistency for me means that file selectors of all needed apps sort files in the same manner, for example).
And, nearly by definition, “ongoing evolutionary-like battles” between different design decisions use to waste resources to some degree.
2004-01-07 1:25 am
Anonymous
IMHO, *that’s* the major flaw with Linux as it’s presented today.”
Huh? It’s the kernel’s fault? IMO, your flaw is the inability to use ie. “man -k” (pardon my French).
that is tuned to specific hardware, and give them the list of the brand names to buy.
I liked some aspects of this idea. It gives the people who have created great systems a good chance to show them off. In showing what they have put together, it may dispell many of the myths of Linux desktops. When people say that a linux does not have a good UI, it can give an example of some great UI’s. When people say that linux multimedia sucks, someone can put together a system that has the kinks in installing jack – http://jackit.sf.net .
What I would like is a chance for people who have created, more or less, roll your own systems, to show them off, and have them considered in future ideas for full-fledged distributions.
What I don’t like so much about this is the emphasis on advertising and splash screens and wallpapers. I have never been a big fan of wallpapers that have the name of the distro or DE on them, I like more personal wallpapers, and sometimes I get sick of those and just set my background to a shade of black or blue.
But the article doesn’t really get to the point of coding things up. I don’t think putting together a really nice system is how it’s done, I think it’s people getting behind an innovative project with very good leadership (such as Linux). What we need most, IMO, is a good collaborative project for GUI’s.
Note that my comment doesn’t make that much sense, in the same way that the story I’m being critical of doesn’t. I’m to lazy to go back and edit it, and that is why you haven’t seen me submit any articles to a computer news site.
but as many are aware, much of that effort is wasted in the way of duplicated work, a great deal of which happen to only be the “sexy” parts of the code base.
————
I wish people would provide references when they say stuff like this. Anybody who keeps up to date on the mailing-lists of various projects knows that this isn’t really true. There are *tons* of people just working behind the scenes, on relatively mundane things. For every guy starting a fork like Xouvert, there are a dozen doing things like making sure fonts with embedded comments work properly (recent Freetype improvement).
The article alluded to tapping into people who couldn’t necessarily code. There was a long thread on kde-usability recently about creating usability groups composed of non-coders. A lot of people thought it would be a good way to work on the “polish” issues with KDE apps, since a lot of that work doesn’t really involve coding. I think the one specific example was one guy whose girlfriend wasn’t a coder, but a big fan of KDE and wanted to contribute. And of course, there are always artists (Everlado!) who contribute their talents.
And I just don’t understand the “duplication of effort” thing. One thing you learn when involved in any sort of engineering work is that there is more than one way to design something. The vast majority of the “duplicated effort” in OSS is not because people don’t want to work with existing projects, but because they want to make different design decisions. That’s why comments along the lines of “merge foo and bar and make something better than both!” always strike me as naive. “foo” and “bar” most likely exist either because they contain mutually exclusive design decisions, or because they are targeted at slightly different markets. You don’t just go and merge such things.
The KDE usability group has existed for quite a while. In my opinion, the biggest thing someone can do to help is build packages for an existing distrobution. The failure rate and dependency issues installing existing packages are probably the largest problem with Linux today. I can’t even count how many times I have gone to install an RPM build for a distro and/or version I am not currently running and had it fail.
In my opinion, the biggest thing someone can do to help is build packages for an existing distrobution
The thing is that OSS often breaks their own compatibility from one version to the other. It is like you clean up the table and then in minutes, someone will mess it again … after a couple of rounds, no sane person will ever want to play the game unless there is money involed.
as far as motherboards/sound cards/video etc (with perhaps the exception of Nvidia) Linux distributions support these very well, especially projects like Knoppix.
the problem is the millions of peripherals which do not conform to any standards (i’m thinking serial/USB) and have no open documentation.
the only time when people will seriously consider using Linux at home is when they can plug all their cheap no-brand peripherals into Linux. this requires a paradigm shift of hardware manufacturers back to the days when they provided specs (not always free, but still) to developers.
These days, a PC without peripherals is, in a way, like a cable TV setop box without signal.
With economy domainated by WalMart and with razor thin profit, you can expect little tech specs
methinks the gates crusher is a bit of a play on words, there ‘s a club in england called gatecrasher: http://www.gatecrasher.co.uk
@ Rainer:
> but as many are aware, much of that effort is wasted in
> the way of duplicated work, a great deal of which happen
> to only be the “sexy” parts of the code base.
> —
> I wish people would provide references when they say
> stuff like this. Anybody who keeps up to date on the
> mailing-lists of various projects knows that this isn’t
> really true.
Take a popular distribution, say SuSE. Do a “complete install” – hard drives are cheap after all, and we don’t want to miss anything. Count the number of programs “designed” to play back MP3.
Where are the different “design decisions”? Unfortunately, the OSS community considers “GUI or not”, “Gnome or KDE” etc. etc. a “design decision”. (In my eyes, a weakness of the Linux UI situation.)
> For every guy starting a fork like Xouvert, there are a
> dozen doing things like making sure fonts with embedded
> comments work properly (recent Freetype improvement).
I seem to miss all the fun stuff… fonts with *embedded* *comments*? What for, ye flippin’ gods? Ups, off-topic…
> And I just don’t understand the “duplication of effort”
> thing. One thing you learn when involved in any sort of
> engineering work is that there is more than one way to
> design something.
Of course. But how many different ways of decoding an MP3 to line-out does one *need*? How many ways of printing a document? How many ways of setting the volume of your sound output?
The problem is probably not so much that these projects *exist*. To the contrary, the more developers the merrier, and to hell with duplicate efforts! We all started with “Hello World!” one day!
But it *is* a problem that the Linux user simply doesn’t start with a “clean” OS and picks his/her choice in tools. They’re free, so they’re all on the CD / the repository, and end up on the user’s hard drive.
It’s fine to have a choice of tools. Hell, the Aminet was the next best thing to sliced bread back in Amiga times, because there was a tool for every taste to be found there.
But the average Linux beginner is dropped right in the middle of a crowded “Start” menu, with a /usr/bin too long to get an overview over, and program names that don’t give a hint as to their purpose. IMHO, *that’s* the major flaw with Linux as it’s presented today.
Yes, there are “build-it-yourself” approaches. Most of them require intricate knowledge of boot process, compiler usage, and kernel architecture.
Just my 0.02€.
Although the name is overblown, this project sounds interesting.
My C coding skills are still at a novice level but I would be glad to help doing documentation.
For all those people who want to “fix” the open source “problems”, please understand why people do what they do.
There are many types of open source projects. Some are commercial and some are recreational.
For commercial projects one has the right to demand polish and attention to all the details one would expect. After all the product is being developed to create income in some form or another. Just like you expect your car’s doors to shut tight when you buy it.
But so many of the software your distribution installs for you are hobby projects. Like when your grandpa makes you a wooden car or your grandma knits you mittens. Depending on their skill level it may be flawless but in my case those things have flaws. If you like what you are given you overlook those flaws because it was a gift after all. You didn’t pay for it. I can just imagine one calling their paternal grandma and saying please don’t give me anymore mittens because my other grandma makes better ones. Both of them do it because it is their hobby and it makes them happy. Hobby software is just the same. It is a gift. It probably is as good as the skill level of the developer(s) and if you don’t like it don’t tell the developers not to make gifts but choose the gifts you do like and put that software in a box on a shelf so to speak.
Until Everyone understands the motives behind a software package these arguments on the “problems” with open source software will be pointless.
If people aren’t careful then developers will stop giving away their handiwork and we will have to pay for all software again.
The reason there is so much “duplication” among open source softwares is because there is a great deal of individual learning going on. There is also a great deal of tweaking and refinement going on which doesn’t always get integrated into the main code tree. When that happens, then naturally there will be a split, often with little different other than a few patches.
The most well known example of “duplicate effort” but wildly different philosophies in implementation, would be KDE and Gnome. Both are desktop environment, there is similar functionality, but the bones of the 2 systems are wildly different. Gnome was started to answer some developers’ qualms over the then non-free Trolltech QT licensing scheme. As long as open source code exists there will be a continual growth of apparently duplicate works but with slight twists, or simply because the author of one software app was unaware of the existance of another similar application written by another. This isn’t a weakness, it’s a strength. Diversity is rarely a weakness.
Another myth about open source, is that “desktop usefullness” is neglected. We have plenty of open source desktop fully functional and featureful applications available. What most open developers *don’t* care about is all the bells and whistles, eye candy, you get with corporate software that is more marketing than useful. The people that created most of the open source software don’t really care if a MS Office dweeb can use their macro on abisuite. We don’t care if a gamer can’t get some weird joystick setup to duplicate in Linux w/ Descent.
Linux and *BSD weren’t developed to replace Microsoft Windows and it’s functionality. They were created to learn computer technology and to have the best damned OS that open developers could create for *ourselves*. They were never meant to be used by morons that can barely figure out how to turn a computer on, let alone how to safely manage software and information security.
Quite frankly, the majority of open developers wouldn’t care one whit should Sun, IBM, etc etc suddenly stop contributing code to the various projects. We’d still use what we created and who cares about the rest of the computing world. Don’t get me wrong, it’s great that they are contributing open technologies to our projects, but I’m not going to worship them for it.
“Take a popular distribution, say SuSE. Do a “complete install” – hard drives are cheap after all, and we don’t want to miss anything. Count the number of programs “designed” to play back MP3.”
Then, do a ldd on them and check out which library they use. That would be either: libmpg321 or libmpg123. So the quality of sound is either 1 or 2. The frontend differs.
And, don’t you think there’s an equal number of these players for Windows? They’re reinventing that wheel too i suppose? It’s just WMP and WA are the most populair. But when you do a full SuSE install you get them all for Free, since they’re Free software? I’d guess that’s the reason.
BTW there’s several aspects why there could be multiple programs. For example some were earlier non-free (Blender?), they were unaware of other programs (does happen..), they like their own design more (Editor-wars), portability (Bochs, Xen, Plex86), student project (Xen), etc etc etc.
You folks that are faulting the article based on the duplicated effort comment, should realize that you are criticizing the editors pre-amble to the article. The author didn’t write that. The actual article starts later.
dr_gonzo wrote: methinks the gates crusher is a bit of a play on words, there ‘s a club in england called gatecrasher: